MediLink Workshop

1
MediLink Workshop 13th November, 2002 Cumann
Ríomheolais Sláinte
2

• What are its objectives?Leveraging the data in
  the EHCR through links to Knowledge.
• Who are involved?
• Researchers mainly in TCD, DIT and SJH.
• Who is funding the work?
• HEA, under its PRTLI Programme

3
Agenda

• Overview
• Electronic Patient Record Server
• Knowledge representation (I)
• Care pathways
• Protocols and guidelines
• Case Based Reasoning
• Knowledge representation (II)
• Models and Simulations
• Demonstration of the MediLink Computing
  Environment

4
Overview

• Some of the challenges
• Ireland does not have nor is it likely to have
  within the short to medium term a regional or
  national electronic record capability.
• This means that for those organisations who do
  wish to make progress some solution to federating
  records needs to be adopted (bringing records and
  parts of records together).
• In addition there needs to be some common method
  by which records can be communicated to multiple
  users.

5
Overview

• Some of the challenges
• Knowledge is most easily represented in a wide
  variety of ways and no one application or method
  can facilitate the capture, utilisation and
  refinement of medical wisdom.
• It follows that some form of framework is needed
  in which many applications can freely
  interoperate to provide a range of services to
  the end-users.

6
Linking Knowledge to Records
Protocols
Workflow
Case-Based Reasoning
Diabetes Modelling
Healthcare Record
Healthcare professionals
Hip Replacement Simulation
7
Linking Knowledge to Records
Protocols
Workflow
Sources of Data
Case-Based Reasoning
Diabetes Modelling
Healthcare Record
Healthcare professionals
Hip Replacement Simulation
8
Communicating with Users (anywhere)
Web Browser
HTTP Client
SynExML
XHTML
Firewall
Servlet Engine
XSLT Processor
XHTML
CORBA
SynExML
SynExML
RSB
CORBA
Feeder
Feeder
9
Workflow - bringing it all together, making it
work !
Guideline Rulebase
Model / Simulation
Semantic Layer
Care Flow Automated Execution Engine
Process Layer
Administrative System
Administrative System
Application Layer

Clinical System
Information Layer
Electronic Healthcare Record
10
Knowledge representation Case Based Reasoning

• Problem solving model based on retrieval of past
  problem episodes
• If we have a case-base of individuals that have
  been labelled as at-risk or not-at-risk we can
  predict the risk status of a new patient by
  finding the nearest neighbours in the case-base.
• Needs descriptions of past cases. These can be
  extracted from patient records (record server).

11
Knowledge representation Guidelines/Protocols
12
Knowledge representation Models
Analyser
LIS
Glucose
Insulin
Electronic Healthcare Record
Insulin Sensitivity calculations carried out
using MINIMAL MODEL method
13
Knowledge representation Simulation
Data from the patient record can be used to
generate loading configurations for a given
patient. Surgical factors can be incorporated by
allowing selection of implant type and
size. Simulation of mechanical and biological
failure processes increase the available
information for decision support.
14
Leveraging Clinical Data

• Use individual patient data to make healthcare
  decision support more effective
• through integrating appropriate tools with
  patient data
• Use collection of patient data (with outcomes) to
  enhance decision support
• enhancing the knowledge bases

15
Knowledge refinement
Patient Records
Knowledge Bases
Decision Tools
Knowledge Refinement

• By internal mechanisms with local data
• And using external sources.

16
Next presentation Electronic Health Records
- an overview including the patient record
server